Runx2 Expression as a Potential Prognostic Marker in Invasive Ductal Breast Carcinoma

The Runx family of transcription factors has been implicated in cancer progression, both positively and negatively. Recent studies assigned a role for Runx2 in promoting breast cancer metastasis. However, the role of Runx2 during the early stage of breast carcinoma and its association with clinical outcomes remain unknown. Assessing the clinicopathological significance of Runx2 expression in a cohort of breast invasive ductal carcinomas (IDC). The correlation of nuclear Runx2 LI with clinicopathological parameters was assessed in 84 IDCs. To study the association of Runx2 with patient outcomes, in addition to treating it as a continuous variable, Runx2 was categorized by its median value (65) and by an additional two cut-off points determined by ROC curve analyses, at 45 for disease free survival (DFS) and 40 for overall survival (OS). Multivariate Cox regression models were also constructed. We used the best subset regression to identify models that predict DFS and OS with as few predictors as possible, and validation was performed. Based on the Predicted R-2, the three best models were identified. Using Cox-regression, the interaction between Runx2 and other clinicopathological terms was tested. Runx2 LI was significantly associated only with positive Her-2 status, and did not correlate significantly with other clinicopathological parameters. Although Runx2 LI, in the continuous form and when categorized by the median, did not correlate significantly with DFS and OS; after it was categorized using the optimal cut-off points determined using ROC curve analysis, the patients with Runx2 LI > 45 % showed a significantly higher event rate and shorter DFS (P = 0.047), whereas patients with Runx2 LI > 40 % showed a significantly shorter OS (P = 0.050). Moreover, Runx2 LI contributed significantly in the models built to predict DFS and OS. For DFS, no interaction terms contributed significantly to the models. However, among stage IV cases, the interaction term between centred Runx2 and ER significantly contributed to the prediction of OS. Runx2 was a significant predictor of OS in this model. Runx2 has a role in biological behaviour and affects the outcome of IDC; therefore, its inhibition may be a new therapeutic strategy. The predictability of Runx2 for OS in stage IV tumours differs with different ER states. The pattern of this difference was not determined because the sample size was not sufficient to allow pattern testing.